#Importing the data and files
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
import seaborn as sns
import seaborn as sns
from scipy import stats; from scipy.stats import zscore, norm, randint

import warnings
warnings.filterwarnings("ignore")

data= pd.read_csv('IHMStefanini_industrial_safety_and_health_database_with_accidents_description.csv')
print("Shape of the dataset is :",data.shape)
data.head()

Shape of the dataset is : (425, 11)

	Unnamed: 0	Data	Countries	Local	Industry Sector	Accident Level	Potential Accident Level	Genre	Employee or Third Party	Critical Risk	Description
0	0	2016-01-01 00:00:00	Country_01	Local_01	Mining	I	IV	Male	Third Party	Pressed	While removing the drill rod of the Jumbo 08 f...
1	1	2016-01-02 00:00:00	Country_02	Local_02	Mining	I	IV	Male	Employee	Pressurized Systems	During the activation of a sodium sulphide pum...
2	2	2016-01-06 00:00:00	Country_01	Local_03	Mining	I	III	Male	Third Party (Remote)	Manual Tools	In the sub-station MILPO located at level +170...
3	3	2016-01-08 00:00:00	Country_01	Local_04	Mining	I	I	Male	Third Party	Others	Being 9:45 am. approximately in the Nv. 1880 C...
4	4	2016-01-10 00:00:00	Country_01	Local_04	Mining	IV	IV	Male	Third Party	Others	Approximately at 11:45 a.m. in circumstances t...

Dropping the unnamed column

data.drop("Unnamed: 0", axis=1, inplace=True)

NullValueCheck

data.isnull().sum()

Data                        0
Countries                   0
Local                       0
Industry Sector             0
Accident Level              0
Potential Accident Level    0
Genre                       0
Employee or Third Party     0
Critical Risk               0
Description                 0
dtype: int64

Duplicate value Check and Deletion

print("Shape of the dataset before duplicates deletion is :",data.shape)
print('Number of duplicates in the dataset :',data.duplicated().sum())
data.drop_duplicates(inplace=True)
print("Shape of the dataset after duplicates deletion is :",data.shape)

Shape of the dataset before duplicates deletion is : (425, 10)
Number of duplicates in the dataset : 7
Shape of the dataset after duplicates deletion is : (418, 10)

print('********Checking the dtypes*********\n')
print(data.dtypes)
print('----------------------------------------------')
print('\n *********Checking the data info********* \n')
print(data.info())

********Checking the dtypes*********

Data                        object
Countries                   object
Local                       object
Industry Sector             object
Accident Level              object
Potential Accident Level    object
Genre                       object
Employee or Third Party     object
Critical Risk               object
Description                 object
dtype: object
----------------------------------------------

 *********Checking the data info********* 

<class 'pandas.core.frame.DataFrame'>
Int64Index: 418 entries, 0 to 424
Data columns (total 10 columns):
 #   Column                    Non-Null Count  Dtype 
---  ------                    --------------  ----- 
 0   Data                      418 non-null    object
 1   Countries                 418 non-null    object
 2   Local                     418 non-null    object
 3   Industry Sector           418 non-null    object
 4   Accident Level            418 non-null    object
 5   Potential Accident Level  418 non-null    object
 6   Genre                     418 non-null    object
 7   Employee or Third Party   418 non-null    object
 8   Critical Risk             418 non-null    object
 9   Description               418 non-null    object
dtypes: object(10)
memory usage: 35.9+ KB
None

From the above, it is clearly evident that all the columns of the data frame are of the type object.

Checking the Statistical Description

data.describe()

	Data	Countries	Local	Industry Sector	Accident Level	Potential Accident Level	Genre	Employee or Third Party	Critical Risk	Description
count	418	418	418	418	418	418	418	418	418	418
unique	287	3	12	3	5	6	2	3	33	411
top	2017-02-08 00:00:00	Country_01	Local_03	Mining	I	IV	Male	Third Party	Others	In the geological reconnaissance activity, in ...
freq	6	248	89	237	309	141	396	185	229	2

#Renaming the data, countries,genre, employee or third party to date, country, gender and nature of employee
data.rename(columns={'Data':'Date', 'Countries':'Country', 'Genre':'Gender', 'Employee or Third Party':'Natureofemployee'}, inplace=True)

data.head(3)

	Date	Country	Local	Industry Sector	Accident Level	Potential Accident Level	Gender	Natureofemployee	Critical Risk	Description
0	2016-01-01 00:00:00	Country_01	Local_01	Mining	I	IV	Male	Third Party	Pressed	While removing the drill rod of the Jumbo 08 f...
1	2016-01-02 00:00:00	Country_02	Local_02	Mining	I	IV	Male	Employee	Pressurized Systems	During the activation of a sodium sulphide pum...
2	2016-01-06 00:00:00	Country_01	Local_03	Mining	I	III	Male	Third Party (Remote)	Manual Tools	In the sub-station MILPO located at level +170...

Finding the unique values of all the columns of the dataframe

col = data[data.columns[~data.columns.isin(['Date','Description'])]].columns.tolist()
for cols in col:
    print(f'Unique values for {cols} is \n{data[cols].unique()}\n')

Unique values for Country is 
['Country_01' 'Country_02' 'Country_03']

Unique values for Local is 
['Local_01' 'Local_02' 'Local_03' 'Local_04' 'Local_05' 'Local_06'
 'Local_07' 'Local_08' 'Local_10' 'Local_09' 'Local_11' 'Local_12']

Unique values for Industry Sector is 
['Mining' 'Metals' 'Others']

Unique values for Accident Level is 
['I' 'IV' 'III' 'II' 'V']

Unique values for Potential Accident Level is 
['IV' 'III' 'I' 'II' 'V' 'VI']

Unique values for Gender is 
['Male' 'Female']

Unique values for Natureofemployee is 
['Third Party' 'Employee' 'Third Party (Remote)']

Unique values for Critical Risk is 
['Pressed' 'Pressurized Systems' 'Manual Tools' 'Others'
 'Fall prevention (same level)' 'Chemical substances' 'Liquid Metal'
 'Electrical installation' 'Confined space'
 'Pressurized Systems / Chemical Substances'
 'Blocking and isolation of energies' 'Suspended Loads' 'Poll' 'Cut'
 'Fall' 'Bees' 'Fall prevention' '\nNot applicable' 'Traffic' 'Projection'
 'Venomous Animals' 'Plates' 'Projection/Burning' 'remains of choco'
 'Vehicles and Mobile Equipment' 'Projection/Choco' 'Machine Protection'
 'Power lock' 'Burn' 'Projection/Manual Tools'
 'Individual protection equipment' 'Electrical Shock'
 'Projection of fragments']

Replacing and Labelling the values of the columns

replace_val = {'Local_01': 1, 'Local_02': 2, 'Local_03': 3, 'Local_04': 4, 'Local_05': 5, 'Local_06': 6, 'Local_07': 7, 'Local_08': 8, 'Local_09': 9, 'Local_10': 10, 'Local_11': 11, 'Local_12': 12}
data['Local'] = data['Local'].map(replace_val)
# replace_val = {'I': 1, 'II': 2, 'III': 3, 'IV': 4, 'V': 5}
# data['Accident Level'] = data['Accident Level'].map(replace_val)
# replace_val = {'I': 0, 'II': 1, 'III': 2, 'IV': 3, 'V': 4, 'VI': 5}
# data['Potential Accident Level'] = data['Potential Accident Level'].map(replace_val)
del replace_val

data.head(5)

	Date	Country	Local	Industry Sector	Accident Level	Potential Accident Level	Gender	Natureofemployee	Critical Risk	Description
0	2016-01-01 00:00:00	Country_01	1	Mining	I	IV	Male	Third Party	Pressed	While removing the drill rod of the Jumbo 08 f...
1	2016-01-02 00:00:00	Country_02	2	Mining	I	IV	Male	Employee	Pressurized Systems	During the activation of a sodium sulphide pum...
2	2016-01-06 00:00:00	Country_01	3	Mining	I	III	Male	Third Party (Remote)	Manual Tools	In the sub-station MILPO located at level +170...
3	2016-01-08 00:00:00	Country_01	4	Mining	I	I	Male	Third Party	Others	Being 9:45 am. approximately in the Nv. 1880 C...
4	2016-01-10 00:00:00	Country_01	4	Mining	IV	IV	Male	Third Party	Others	Approximately at 11:45 a.m. in circumstances t...

Changing the date column to date format

data['Date'] = pd.to_datetime(data['Date'])

Extracting the month, year and day column to the data.

data['Year'] = data['Date'].apply(lambda x : x.year)
data['Month'] = data['Date'].apply(lambda x : x.month)
data['Weekday'] = data['Date'].apply(lambda x : x.day_name())

data.head()

	Date	Country	Local	Industry Sector	Accident Level	Potential Accident Level	Gender	Natureofemployee	Critical Risk	Description	Year	Month	Weekday
0	2016-01-01	Country_01	1	Mining	I	IV	Male	Third Party	Pressed	While removing the drill rod of the Jumbo 08 f...	2016	1	Friday
1	2016-01-02	Country_02	2	Mining	I	IV	Male	Employee	Pressurized Systems	During the activation of a sodium sulphide pum...	2016	1	Saturday
2	2016-01-06	Country_01	3	Mining	I	III	Male	Third Party (Remote)	Manual Tools	In the sub-station MILPO located at level +170...	2016	1	Wednesday
3	2016-01-08	Country_01	4	Mining	I	I	Male	Third Party	Others	Being 9:45 am. approximately in the Nv. 1880 C...	2016	1	Friday
4	2016-01-10	Country_01	4	Mining	IV	IV	Male	Third Party	Others	Approximately at 11:45 a.m. in circumstances t...	2016	1	Sunday

Function to create month into seasons

# function to create month into seasons
def convert_to_season(x):
    if x in [9, 10, 11]:
        season = 'Spring'
    elif x in [12, 1, 2]:
        season = 'Summer'
    elif x in [3, 4, 5]:
        season = 'Autumn'
    elif x in [6, 7, 8]:
        season = 'Winter'
    return season

data['Season'] = data['Month'].apply(convert_to_season)
data.head(3)

	Date	Country	Local	Industry Sector	Accident Level	Potential Accident Level	Gender	Natureofemployee	Critical Risk	Description	Year	Month	Weekday	Season
0	2016-01-01	Country_01	1	Mining	I	IV	Male	Third Party	Pressed	While removing the drill rod of the Jumbo 08 f...	2016	1	Friday	Summer
1	2016-01-02	Country_02	2	Mining	I	IV	Male	Employee	Pressurized Systems	During the activation of a sodium sulphide pum...	2016	1	Saturday	Summer
2	2016-01-06	Country_01	3	Mining	I	III	Male	Third Party (Remote)	Manual Tools	In the sub-station MILPO located at level +170...	2016	1	Wednesday	Summer

Univariate analysis covering Count, Pie, Scatter, Histogram plots

features=['Country', 'Local', 'Industry Sector', 'Accident Level',
       'Potential Accident Level', 'Gender', 'Natureofemployee',
       'Critical Risk', 'Year', 'Month', 'Weekday', 'Season']

def univariate_analysis_categorical(dataset,feature):    
    print("\n")
    print("===========================================================================================")
    print("Univariate Analysis of feature: ",feature)
    print("===========================================================================================\n")
        
           
    print("Unique values: ",feature)
    print("-----------------")        
    print(dataset[feature].unique()) 
        
        
    print("\n")  
    print("-----------------")
    print("Countplot  for feature: ",feature)
    print("-----------------")
        
    plt.figure(figsize=(10,6))
    sns.countplot(dataset[feature],order = dataset[feature].value_counts().index)
    plt.xticks(rotation = 'vertical')
    plt.show()
    
    print("-----------------")
    print("Pie Chart for feature: ",feature)
    print("------------------")      
        
    labels=dataset[feature].unique()
    plt.figure(figsize=(10,6))
    dataset[feature].value_counts().plot.pie(autopct="%.1f%%")
    plt.show()
        
    print("\n")
    print("-----------------")
    print("Histplot  for feature: ",feature)
    print("-------------------")
        
    plt.figure(figsize=(10,6))
    sns.histplot(dataset[feature])
    plt.show()      
    
    print("\n")
    print("-----------------")
    print("Value Counts for feature: ",feature)
    print("-------------------")
    
    print(dataset[feature].value_counts().sort_values(ascending=False))

#!pip install pyqt5

1.Country

univariate_analysis_categorical(data,'Country')

===========================================================================================
Univariate Analysis of feature:  Country
===========================================================================================

Unique values:  Country
-----------------
['Country_01' 'Country_02' 'Country_03']


-----------------
Countplot  for feature:  Country
-----------------

-----------------
Pie Chart for feature:  Country
------------------

-----------------
Histplot  for feature:  Country
-------------------

-----------------
Value Counts for feature:  Country
-------------------
Country_01    248
Country_02    129
Country_03     41
Name: Country, dtype: int64

----- From the above plots, we can conclude the following

1. The country_01 has a count of about 248. Country _02 has a count of about 129. Country_03 has a count of about 41.

2. From the above pie chart, it can be infered that the country _01 is the most affected country with about 59% accidents and country_03 is the least affected country.

3. From the above output, the country_01 has maximum accidents and country_03 has minimum accidents.

2.Local

#Count plot
univariate_analysis_categorical(data,'Local')

===========================================================================================
Univariate Analysis of feature:  Local
===========================================================================================

Unique values:  Local
-----------------
[ 1  2  3  4  5  6  7  8 10  9 11 12]


-----------------
Countplot  for feature:  Local
-----------------

-----------------
Pie Chart for feature:  Local
------------------

-----------------
Histplot  for feature:  Local
-------------------

-----------------
Value Counts for feature:  Local
-------------------
3     89
5     59
1     56
4     55
6     46
10    41
8     27
2     23
7     14
12     4
9      2
11     2
Name: Local, dtype: int64

1. Most of the accidents happen in Local_03 with a count of about 89.
2. Maximum accidents are taken place in local_03 with 21.18% and least accidents are taken place in local_09 and local_11 with 0.47%.
3. From the above output, it can be infered that the local_03 is more prone to accidents.
4. From the above histogram, it can be observed that the number of accidents in Local_03 are about 90.

3.Industry sector

univariate_analysis_categorical(data,'Industry Sector')

===========================================================================================
Univariate Analysis of feature:  Industry Sector
===========================================================================================

Unique values:  Industry Sector
-----------------
['Mining' 'Metals' 'Others']


-----------------
Countplot  for feature:  Industry Sector
-----------------

-----------------
Pie Chart for feature:  Industry Sector
------------------

-----------------
Histplot  for feature:  Industry Sector
-------------------

-----------------
Value Counts for feature:  Industry Sector
-------------------
Mining    237
Metals    134
Others     47
Name: Industry Sector, dtype: int64

1. From the above, it is evident that the mining is prone to more accidents with about 237.
2. It can be observed that the mostly affected sector is Mining sector. 56.71% of accidents occur in Mining sector.
3. It is clearly evident that the mining is prone to more accidents

4.Accident level

univariate_analysis_categorical(data,'Accident Level')

===========================================================================================
Univariate Analysis of feature:  Accident Level
===========================================================================================

Unique values:  Accident Level
-----------------
['I' 'IV' 'III' 'II' 'V']


-----------------
Countplot  for feature:  Accident Level
-----------------

-----------------
Pie Chart for feature:  Accident Level
------------------

-----------------
Histplot  for feature:  Accident Level
-------------------

-----------------
Value Counts for feature:  Accident Level
-------------------
I      309
II      40
III     31
IV      30
V        8
Name: Accident Level, dtype: int64

1. From the above count plot, it is clearly evident that the most accidents belongs to "Accident Level" "1" with a count of about more than 300.
2. From the above pie chart, it can be determined that the maximum accidents are of level 1 equivalent to about 73.9%

5.potential accident level

univariate_analysis_categorical(data,'Potential Accident Level')

===========================================================================================
Univariate Analysis of feature:  Potential Accident Level
===========================================================================================

Unique values:  Potential Accident Level
-----------------
['IV' 'III' 'I' 'II' 'V' 'VI']


-----------------
Countplot  for feature:  Potential Accident Level
-----------------

-----------------
Pie Chart for feature:  Potential Accident Level
------------------

-----------------
Histplot  for feature:  Potential Accident Level
-------------------

-----------------
Value Counts for feature:  Potential Accident Level
-------------------
IV     141
III    106
II      95
I       45
V       30
VI       1
Name: Potential Accident Level, dtype: int64

1. From the above count plot, it can be determined that the most "Potential Accident Level" belongs to level IV with a count of about 141.
2. From the above pie chart, it is evident that most "Potential Accident Level" belongs to level IV with 33.7%.

6.Gender

univariate_analysis_categorical(data,'Gender')

===========================================================================================
Univariate Analysis of feature:  Gender
===========================================================================================

Unique values:  Gender
-----------------
['Male' 'Female']


-----------------
Countplot  for feature:  Gender
-----------------

-----------------
Pie Chart for feature:  Gender
------------------

-----------------
Histplot  for feature:  Gender
-------------------

-----------------
Value Counts for feature:  Gender
-------------------
Male      396
Female     22
Name: Gender, dtype: int64

1. From the above plot, it is evident that Most affected wokers in accidents are male with a count of 396.
2. From the above pie chart, it is evient that most affected wokers in accidents are male.

7.Nature of the employee

univariate_analysis_categorical(data,'Natureofemployee')

===========================================================================================
Univariate Analysis of feature:  Natureofemployee
===========================================================================================

Unique values:  Natureofemployee
-----------------
['Third Party' 'Employee' 'Third Party (Remote)']


-----------------
Countplot  for feature:  Natureofemployee
-----------------

-----------------
Pie Chart for feature:  Natureofemployee
------------------

-----------------
Histplot  for feature:  Natureofemployee
-------------------

-----------------
Value Counts for feature:  Natureofemployee
-------------------
Third Party             185
Employee                178
Third Party (Remote)     55
Name: Natureofemployee, dtype: int64

From the above it can be determined that the employee type of Third party are prone to accidents.

8.Critical risk

#Count plot
# plt.figure(figsize=(20,5))
# descending_order = data['Critical Risk'].value_counts().sort_values(ascending=False).index
# sns.countplot(x=data['Critical Risk'],order=descending_order);
# plt.xticks(rotation = 'vertical')

univariate_analysis_categorical(data,'Critical Risk')

===========================================================================================
Univariate Analysis of feature:  Critical Risk
===========================================================================================

Unique values:  Critical Risk
-----------------
['Pressed' 'Pressurized Systems' 'Manual Tools' 'Others'
 'Fall prevention (same level)' 'Chemical substances' 'Liquid Metal'
 'Electrical installation' 'Confined space'
 'Pressurized Systems / Chemical Substances'
 'Blocking and isolation of energies' 'Suspended Loads' 'Poll' 'Cut'
 'Fall' 'Bees' 'Fall prevention' '\nNot applicable' 'Traffic' 'Projection'
 'Venomous Animals' 'Plates' 'Projection/Burning' 'remains of choco'
 'Vehicles and Mobile Equipment' 'Projection/Choco' 'Machine Protection'
 'Power lock' 'Burn' 'Projection/Manual Tools'
 'Individual protection equipment' 'Electrical Shock'
 'Projection of fragments']


-----------------
Countplot  for feature:  Critical Risk
-----------------

-----------------
Pie Chart for feature:  Critical Risk
------------------

-----------------
Histplot  for feature:  Critical Risk
-------------------

-----------------
Value Counts for feature:  Critical Risk
-------------------
Others                                       229
Pressed                                       24
Manual Tools                                  20
Chemical substances                           17
Cut                                           14
Projection                                    13
Venomous Animals                              13
Bees                                          10
Fall                                           9
Vehicles and Mobile Equipment                  8
remains of choco                               7
Fall prevention (same level)                   7
Pressurized Systems                            7
Suspended Loads                                6
Fall prevention                                6
Pressurized Systems / Chemical Substances      3
Power lock                                     3
Blocking and isolation of energies             3
Liquid Metal                                   3
Electrical Shock                               2
Machine Protection                             2
Projection/Choco                               1
Burn                                           1
Projection/Manual Tools                        1
Electrical installation                        1
\nNot applicable                               1
Poll                                           1
Projection/Burning                             1
Plates                                         1
Individual protection equipment                1
Traffic                                        1
Projection of fragments                        1
Confined space                                 1
Name: Critical Risk, dtype: int64

When we count the number of incidents by each type of critical risk, Others tops the list.

9.Year

univariate_analysis_categorical(data,'Year')

===========================================================================================
Univariate Analysis of feature:  Year
===========================================================================================

Unique values:  Year
-----------------
[2016 2017]


-----------------
Countplot  for feature:  Year
-----------------

-----------------
Pie Chart for feature:  Year
------------------

-----------------
Histplot  for feature:  Year
-------------------

-----------------
Value Counts for feature:  Year
-------------------
2016    283
2017    135
Name: Year, dtype: int64

From the above, it is clearly evident that most accidents happend in year 2016. i.e- more than 250.

10.Month

univariate_analysis_categorical(data,'Month')

===========================================================================================
Univariate Analysis of feature:  Month
===========================================================================================

Unique values:  Month
-----------------
[ 1  2  3  4  5  6  7  8  9 10 11 12]


-----------------
Countplot  for feature:  Month
-----------------

-----------------
Pie Chart for feature:  Month
------------------

-----------------
Histplot  for feature:  Month
-------------------

-----------------
Value Counts for feature:  Month
-------------------
2     61
4     51
6     51
3     50
5     40
1     39
7     24
9     24
12    23
8     21
10    21
11    13
Name: Month, dtype: int64

1. From the above, it can be determined that the most accidents are of month Feb.
2. From the above, it is evident that most of the accidents happened in feb equivalent to 14.6%.

11.WeekDay

univariate_analysis_categorical(data,'Weekday')

===========================================================================================
Univariate Analysis of feature:  Weekday
===========================================================================================

Unique values:  Weekday
-----------------
['Friday' 'Saturday' 'Wednesday' 'Sunday' 'Tuesday' 'Thursday' 'Monday']


-----------------
Countplot  for feature:  Weekday
-----------------

-----------------
Pie Chart for feature:  Weekday
------------------

-----------------
Histplot  for feature:  Weekday
-------------------

-----------------
Value Counts for feature:  Weekday
-------------------
Thursday     76
Tuesday      69
Wednesday    62
Friday       61
Saturday     56
Monday       53
Sunday       41
Name: Weekday, dtype: int64

1. From the above count plot, it can be determined that max accidents happened on thursday with approximately 76 accidents.
2. From the above piechart, it is evident that most accidents happend in Thursday equivalent to 18.2%

Bivariate analysis

Gender vs RestAll

1.Gender vs Accident level

sns.countplot(x="Accident Level",hue="Gender", data=data)

<AxesSubplot:xlabel='Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Accident Level'],columns=data['Gender'])

print("------------------------------------------")
print("Cross table Analysis of features: ",'Accident Level',' and  ', 'Gender')
print("------------------------------------------")
display(bivariate_analysis_df)  

------------------------------------------
Cross table Analysis of features:  Accident Level  and   Gender
------------------------------------------

Gender	Female	Male
Accident Level
I	18	291
II	3	37
III	1	30
IV	0	30
V	0	8

From the above count plot, it can be determined that the most of the accidents happened at level I with gender male.

2.Gender vs Potential Accident Level

sns.countplot(x="Potential Accident Level",hue="Gender", data=data)

<AxesSubplot:xlabel='Potential Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Potential Accident Level',' and  ', 'Gender')
print("--------------------------------------------------------------------------")
display(bivariate_analysis_df)  

 Cross table Analysis of features:  Potential Accident Level  and   Gender
--------------------------------------------------------------------------

Gender	Female	Male
Potential Accident Level
I	0	45
II	14	81
III	3	103
IV	4	137
V	1	29
VI	0	1

From the above,it can be determined that most of the potential level accidents happened to male compared to female, of which Potential Accident Level of IV is dominant

3.Gender vs Country

sns.countplot(x="Country",hue="Gender", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df)  

 Cross table Analysis of features:  Country  and   Gender
------------------------------------------------------

Gender	Female	Male
Country
Country_01	7	241
Country_02	15	114
Country_03	0	41

From the above countplot, it can be determined that the maximum number of accidents took place in country_01 to males and they are about 241.

4.Gender vs Industry Sector

#count plot to determine the number of accidents happened due to industry sector with their gender
sns.countplot(x="Industry Sector",hue="Gender", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df)  

 Cross table Analysis of features:  Industry Sector  and   Gender
------------------------------------------------------

Gender	Female	Male
Industry Sector
Metals	13	121
Mining	5	232
Others	4	43

From the above count plot, it is evident that most of the accidents happened to Male in the mining sector, around 232.

5.Gender vs Year

#Countplot to find in which max accidents took place to both female and male
sns.countplot(x="Year",hue="Gender", data=data)

<AxesSubplot:xlabel='Year', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Year'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Year',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df)  

 Cross table Analysis of features:  Year  and   Gender
------------------------------------------------------

Gender	Female	Male
Year
2016	14	269
2017	8	127

From the above countplot, it is clearly evident that maximum accidents took place in 2016 to the male when compared to female with a count of 269.

6.Gender vs month

#Countplot to determine in which month the maximum accidents took place to both female and males
sns.countplot(x="Month",hue="Gender", data=data)

<AxesSubplot:xlabel='Month', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Month'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Month',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df)  

 Cross table Analysis of features:  Month  and   Gender
------------------------------------------------------

Gender	Female	Male
Month
1	0	39
2	4	57
3	1	49
4	1	50
5	4	36
6	1	50
7	4	20
8	3	18
9	3	21
10	0	21
11	0	13
12	1	22

From the above count plot, it is determined that maximum number of accidents happened to male in the month feb with a count 57.

7.Gender vs weekday

#Countplot to find out on which day max accidents took place to the both genders
sns.countplot(x="Weekday",hue="Gender", data=data)

<AxesSubplot:xlabel='Weekday', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Weekday'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Weekday',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Weekday  and   Gender
------------------------------------------------------

Gender	Female	Male
Weekday
Friday	2	59
Monday	4	49
Saturday	1	55
Sunday	2	39
Thursday	3	73
Tuesday	6	63
Wednesday	4	58

Max accidents happened to male on thursday with a count of more than 73

8. Gender vs Nature of Employee

#count plot to determine which type of employee and gender faced most of the accidents
sns.countplot(x="Natureofemployee",hue="Gender", data=data)

<AxesSubplot:xlabel='Natureofemployee', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Natureofemployee'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Natureofemployee',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Natureofemployee  and   Gender
------------------------------------------------------

Gender	Female	Male
Natureofemployee
Employee	8	170
Third Party	9	176
Third Party (Remote)	5	50

From the above output, it is clearly evident that maximum accidents happened to third party male employees. i.e- 176.

9. Gender vs Critical Risk

#count plot to determine which type of employee and gender faced most of the accidents
sns.countplot(x="Critical Risk",hue="Gender", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)
plt.xticks(rotation=90)

(array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]),
 [Text(0, 0, 'Pressed'),
  Text(1, 0, 'Pressurized Systems'),
  Text(2, 0, 'Manual Tools'),
  Text(3, 0, 'Others'),
  Text(4, 0, 'Fall prevention (same level)'),
  Text(5, 0, 'Chemical substances'),
  Text(6, 0, 'Liquid Metal'),
  Text(7, 0, 'Electrical installation'),
  Text(8, 0, 'Confined space'),
  Text(9, 0, 'Pressurized Systems / Chemical Substances'),
  Text(10, 0, 'Blocking and isolation of energies'),
  Text(11, 0, 'Suspended Loads'),
  Text(12, 0, 'Poll'),
  Text(13, 0, 'Cut'),
  Text(14, 0, 'Fall'),
  Text(15, 0, 'Bees'),
  Text(16, 0, 'Fall prevention'),
  Text(17, 0, '\nNot applicable'),
  Text(18, 0, 'Traffic'),
  Text(19, 0, 'Projection'),
  Text(20, 0, 'Venomous Animals'),
  Text(21, 0, 'Plates'),
  Text(22, 0, 'Projection/Burning'),
  Text(23, 0, 'remains of choco'),
  Text(24, 0, 'Vehicles and Mobile Equipment'),
  Text(25, 0, 'Projection/Choco'),
  Text(26, 0, 'Machine Protection'),
  Text(27, 0, 'Power lock'),
  Text(28, 0, 'Burn'),
  Text(29, 0, 'Projection/Manual Tools'),
  Text(30, 0, 'Individual protection equipment'),
  Text(31, 0, 'Electrical Shock'),
  Text(32, 0, 'Projection of fragments')])

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Gender'])

print("\n Cross table Analysis of features: ",'Critical Risk',' and  ', 'Gender')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Critical Risk  and   Gender
------------------------------------------------------

Gender	Female	Male
Critical Risk
\nNot applicable	0	1
Bees	0	10
Blocking and isolation of energies	0	3
Burn	0	1
Chemical substances	4	13
Confined space	0	1
Cut	3	11
Electrical Shock	0	2
Electrical installation	0	1
Fall	1	8
Fall prevention	0	6
Fall prevention (same level)	1	6
Individual protection equipment	0	1
Liquid Metal	0	3
Machine Protection	0	2
Manual Tools	1	19
Others	9	220
Plates	0	1
Poll	0	1
Power lock	0	3
Pressed	0	24
Pressurized Systems	1	6
Pressurized Systems / Chemical Substances	0	3
Projection	0	13
Projection of fragments	0	1
Projection/Burning	0	1
Projection/Choco	0	1
Projection/Manual Tools	0	1
Suspended Loads	0	6
Traffic	1	0
Vehicles and Mobile Equipment	0	8
Venomous Animals	1	12
remains of choco	0	7

Critical Risk of type "Others" is dominant across both Male and Female Genders

Industry sector vs RestAll

1. Industry Sector Vs Accident Level

#Countplot to determine the accident level happened at different industry sectors
sns.countplot(x="Industry Sector",hue="Accident Level", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Accident Level'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Accident Level
------------------------------------------------------

Accident Level	I	II	III	IV	V
Industry Sector
Metals	107	12	7	7	1
Mining	163	26	20	21	7
Others	39	2	4	2	0

Maximum number of accidents happened in the mining sector with accident Level I. i.e- 163.

2.Industry sector vs potential accident level

#Countplot to determine the potential accident level according to the industry sector
sns.countplot(x="Industry Sector",hue="Potential Accident Level", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Potential Accident Level'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Potential Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Potential Accident Level
------------------------------------------------------

Potential Accident Level	I	II	III	IV	V	VI
Industry Sector
Metals	7	48	44	33	2	0
Mining	8	40	61	99	28	1
Others	30	7	1	9	0	0

Maximum number of accidents happened in the potential accident level 4 and mining sector with a count 99. Minimum number of accidents took place in the mining sector at a potential accident level 6.

3.Industry Sector vs Critical Risk

#Countplot to determine the number of accidents taken place at the industry sector wrt critical risk
fig = plt.figure(figsize = (15, 7.2))
ax = fig.add_subplot(121)
sns.countplot(x = 'Critical Risk', data = data, ax = ax, orient = 'v',
                  hue = 'Industry Sector')
plt.legend(labels = data['Industry Sector'].unique())
plt.xticks(rotation = 90)

(array([ 0,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 15, 16,
        17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32]),
 [Text(0, 0, 'Pressed'),
  Text(1, 0, 'Pressurized Systems'),
  Text(2, 0, 'Manual Tools'),
  Text(3, 0, 'Others'),
  Text(4, 0, 'Fall prevention (same level)'),
  Text(5, 0, 'Chemical substances'),
  Text(6, 0, 'Liquid Metal'),
  Text(7, 0, 'Electrical installation'),
  Text(8, 0, 'Confined space'),
  Text(9, 0, 'Pressurized Systems / Chemical Substances'),
  Text(10, 0, 'Blocking and isolation of energies'),
  Text(11, 0, 'Suspended Loads'),
  Text(12, 0, 'Poll'),
  Text(13, 0, 'Cut'),
  Text(14, 0, 'Fall'),
  Text(15, 0, 'Bees'),
  Text(16, 0, 'Fall prevention'),
  Text(17, 0, '\nNot applicable'),
  Text(18, 0, 'Traffic'),
  Text(19, 0, 'Projection'),
  Text(20, 0, 'Venomous Animals'),
  Text(21, 0, 'Plates'),
  Text(22, 0, 'Projection/Burning'),
  Text(23, 0, 'remains of choco'),
  Text(24, 0, 'Vehicles and Mobile Equipment'),
  Text(25, 0, 'Projection/Choco'),
  Text(26, 0, 'Machine Protection'),
  Text(27, 0, 'Power lock'),
  Text(28, 0, 'Burn'),
  Text(29, 0, 'Projection/Manual Tools'),
  Text(30, 0, 'Individual protection equipment'),
  Text(31, 0, 'Electrical Shock'),
  Text(32, 0, 'Projection of fragments')])

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Industry Sector'])

print("\n Cross table Analysis of features: ",'Critical Risk',' and  ', 'Industry Sector')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Critical Risk  and   Industry Sector
------------------------------------------------------

Industry Sector	Metals	Mining	Others
Critical Risk
\nNot applicable	1	0	0
Bees	0	0	10
Blocking and isolation of energies	3	0	0
Burn	1	0	0
Chemical substances	15	2	0
Confined space	1	0	0
Cut	10	4	0
Electrical Shock	0	2	0
Electrical installation	0	1	0
Fall	2	5	2
Fall prevention	3	2	1
Fall prevention (same level)	6	1	0
Individual protection equipment	0	1	0
Liquid Metal	3	0	0
Machine Protection	2	0	0
Manual Tools	14	5	1
Others	33	176	20
Plates	1	0	0
Poll	0	0	1
Power lock	1	2	0
Pressed	17	7	0
Pressurized Systems	6	1	0
Pressurized Systems / Chemical Substances	3	0	0
Projection	4	9	0
Projection of fragments	0	1	0
Projection/Burning	1	0	0
Projection/Choco	0	0	1
Projection/Manual Tools	0	1	0
Suspended Loads	5	1	0
Traffic	0	0	1
Vehicles and Mobile Equipment	0	8	0
Venomous Animals	2	1	10
remains of choco	0	7	0

From the above count plot, it is evident that maximum number of accidents happened in mining with a critical risk of others. i.e- about 175

4.Industry sector vs Local

sns.countplot(x="Local",hue="Industry Sector", data=data)

<AxesSubplot:xlabel='Local', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Local')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Local
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Industry Sector
Metals	0	0	0	0	59	46	0	27	2	0	0	0
Mining	56	23	89	55	0	0	14	0	0	0	0	0
Others	0	0	0	0	0	0	0	0	0	41	2	4

Many accidents happened with a local 3 and industrial sector mining. i.e- more than 80. Least accidents took place with local 11 and industrial sector others.

5. Industry Sector Vs Year

#Count plot to determine the number of accidents taken place in year 2016 and 2017 according to the industrial sector
sns.countplot(x="Year",hue="Industry Sector", data=data)

<AxesSubplot:xlabel='Year', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Year'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Year
------------------------------------------------------

Year	2016	2017
Industry Sector
Metals	97	37
Mining	159	78
Others	27	20

From the above plot, the following could be determined

1.The number of accidents taken place in year 2016 for mining sector is 160.

2.The number of accidents taken place in year 2016 wrt metals sector is about 100.

3.The number of accidents taken place in the year 2016 wrt others sector is about 30. Hence, it can be determined that maximum accidents took place in mining sector in the year 2016.

4.The number of accidents taken place in the year 2017 wrt mining sector is 80.

5.The number of accidents taken place in the year 2017 wrt metals sector is about 40.

6.The number of accidents taken place in the year 2017 wrt others sector is 20.

Hence, it can be determined that max accidents took place in mining sector in the year 2017

6. Industry Sector Vs Month

#Count plot to determine the accidents taken place in all the months wrt industrial sector
sns.countplot(x="Industry Sector",hue="Month", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a5073e6f40>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Month'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Month')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Month
------------------------------------------------------

Month	1	2	3	4	5	6	7	8	9	10	11	12
Industry Sector
Metals	13	19	17	15	12	11	9	9	11	5	5	8
Mining	24	38	26	31	20	28	13	10	10	15	7	15
Others	2	4	7	5	8	12	2	2	3	1	1	0

Maximum number of accidents happened in the month feb and mining sector. The least number of accidents took place in the others sector and month december.

7.Industry sector vs Weekday

sns.countplot(x="Industry Sector",hue="Weekday", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Weekday'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Weekday')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Weekday
------------------------------------------------------

Weekday	Friday	Monday	Saturday	Sunday	Thursday	Tuesday	Wednesday
Industry Sector
Metals	20	24	6	15	25	21	23
Mining	40	26	44	25	37	36	29
Others	1	3	6	1	14	12	10

Maximum number of accidents hapenned on the day saturday in the mining sector. i.e- more than 40. The least number of accidents happened on the day sunday in the others sector.

8. Industry sector Vs country

sns.countplot(x="Industry Sector",hue="Country", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Country'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Country')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Country
------------------------------------------------------

Country	Country_01	Country_02	Country_03
Industry Sector
Metals	46	88	0
Mining	200	37	0
Others	2	4	41

From the above count plot, it is evident that the maximum number of accidents took place in country_01 and mining sector.i.e- 200. The least number of accidents took place in country _01 and others sector.

9.Industry Sector Vs nature of employee

sns.countplot(x="Industry Sector",hue="Natureofemployee", data=data)

<AxesSubplot:xlabel='Industry Sector', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Industry Sector'],columns=data['Natureofemployee'])

print("\n Cross table Analysis of features: ",'Industry Sector',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Industry Sector  and   Natureofemployee
------------------------------------------------------

Natureofemployee	Employee	Third Party	Third Party (Remote)
Industry Sector
Metals	76	31	27
Mining	89	120	28
Others	13	34	0

From the above count plot, it is clearly evident that the maximum accidents took place in the mining sector with the third party employee type. i.e- about 120. The least number of accidents took place in the others sectors with the nature of employee as employee.

Country vs RestAll

1. Country vs Year

sns.countplot(x="Country",hue="Year", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Year'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Year
------------------------------------------------------

Year	2016	2017
Country
Country_01	174	74
Country_02	86	43
Country_03	23	18

From the above output, the following can be determined-

1.The number of accidents taken place in country_01 and year 2016 is 174.

2.The number of accidents taken place in country_01 and year 2017 is about 74.

3.The number of accidents taken place in country_02 and year 2016 is more than 86.

4.The number of accidents taken place in country_02 and year 2017 is about 43.

5.The number of accidents taken place in country_03 and year 2016 is about 23.

6.The number of accidents taken place in country_03 and year 2017 is about 18.

2. Country Vs accident level

sns.countplot(x="Country",hue="Accident Level", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Accident Level'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Accident Level
------------------------------------------------------

Accident Level	I	II	III	IV	V
Country
Country_01	177	19	21	23	8
Country_02	98	19	7	5	0
Country_03	34	2	3	2	0

From the above count plot, it is clearly evident that the maximum number of accidents took place in accident level 1 and country_01.

3. Country Vs Potential Accident Level

sns.countplot(x="Country",hue="Potential Accident Level", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Potential Accident Level'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Potential Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Potential Accident Level
------------------------------------------------------

Potential Accident Level	I	II	III	IV	V	VI
Country
Country_01	10	51	64	101	21	1
Country_02	6	40	41	33	9	0
Country_03	29	4	1	7	0	0

From the above plot, it is evident that the maximum accidents occurred in country_01 and potential accident level 3.

4. Country Vs Local

sns.countplot(x="Country",hue="Local", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a507915bb0>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Local')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Local
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Country
Country_01	56	0	89	55	0	46	0	0	0	0	2	0
Country_02	0	23	0	0	59	0	14	27	2	0	0	4
Country_03	0	0	0	0	0	0	0	0	0	41	0	0

Country 1 is more dominant in local 3 region and least dominant in Local 12

5.Country Vs Nature Of Employee

sns.countplot(x="Country",hue="Natureofemployee", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Country'],columns=data['Natureofemployee'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Natureofemployee
------------------------------------------------------

Natureofemployee	Employee	Third Party	Third Party (Remote)
Country
Country_01	87	138	23
Country_02	84	13	32
Country_03	7	34	0

Accidents in Country 01 is more dominant in Third Party type of employee, country 03 is least dominant in Third Party (Remote)

6. Country Vs Critical Risk

sns.countplot(x="Country",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a507a6ec10>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Country'])

print("\n Cross table Analysis of features: ",'Country',' and  ', 'Critical Risk')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Country  and   Critical Risk
------------------------------------------------------

Country	Country_01	Country_02	Country_03
Critical Risk
\nNot applicable	0	1	0
Bees	0	0	10
Blocking and isolation of energies	1	2	0
Burn	0	1	0
Chemical substances	4	13	0
Confined space	0	1	0
Cut	5	9	0
Electrical Shock	2	0	0
Electrical installation	1	0	0
Fall	6	1	2
Fall prevention	3	2	1
Fall prevention (same level)	5	2	0
Individual protection equipment	1	0	0
Liquid Metal	0	3	0
Machine Protection	1	1	0
Manual Tools	7	12	1
Others	169	45	15
Plates	1	0	0
Poll	0	0	1
Power lock	3	0	0
Pressed	9	15	0
Pressurized Systems	1	6	0
Pressurized Systems / Chemical Substances	2	1	0
Projection	9	4	0
Projection of fragments	1	0	0
Projection/Burning	0	1	0
Projection/Choco	0	0	1
Projection/Manual Tools	1	0	0
Suspended Loads	3	3	0
Traffic	0	1	0
Vehicles and Mobile Equipment	7	1	0
Venomous Animals	0	3	10
remains of choco	6	1	0

Country 01 is more dominant in Others Critical Risk and Critical Risk is least dominant in Country 03

Local Vs Rest All

1. Local Vs Accident Level

sns.countplot(x="Local",hue="Accident Level", data=data)

<AxesSubplot:xlabel='Local', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Accident Level'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Local  and   Accident Level
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Accident Level
I	45	14	65	30	51	36	9	19	1	34	1	4
II	1	6	8	9	6	1	2	5	0	2	0	0
III	5	2	5	7	2	3	1	1	1	3	1	0
IV	4	1	8	6	0	5	2	2	0	2	0	0
V	1	0	3	3	0	1	0	0	0	0	0	0

Accident level 1 is more dominant in Local 2 region with 65 accidents, while Accident Level V is least across all Locals

2. Local Vs Potential Accident Level

sns.countplot(x="Local",hue="Potential Accident Level", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a5077f0100>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Potential Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df) 

 Cross table Analysis of features:  Local  and   Potential Accident Level
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Potential Accident Level
I	2	1	2	2	1	4	1	2	0	29	0	1
II	7	6	12	11	15	20	4	12	1	4	1	2
III	13	6	21	19	25	11	2	8	0	1	0	0
IV	30	7	41	19	18	10	2	4	1	7	1	1
V	4	3	12	4	0	1	5	1	0	0	0	0
VI	0	0	1	0	0	0	0	0	0	0	0	0

Overall Local 3 is more prone to Multiple potential accidents, while local 12 is the least

3. Local Vs Natureofemployee

sns.countplot(x="Local",hue="Natureofemployee", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a507faeeb0>

bivariate_analysis_df = pd.crosstab(index=data['Natureofemployee'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Local  and   Natureofemployee
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Natureofemployee
Employee	23	11	30	14	37	18	11	19	2	7	2	4
Third Party	30	1	48	40	9	20	1	2	0	34	0	0
Third Party (Remote)	3	11	11	1	13	8	2	6	0	0	0	0

Type Employee is more dominant across all Locals, while Type Third Party(Remote) is least dominant across all Locals

4. Local Vs Critical Risk

sns.countplot(x="Local",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a50802b970>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Critical Risk')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Local  and   Critical Risk
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Critical Risk
\nNot applicable	0	0	0	0	0	0	0	1	0	0	0	0
Bees	0	0	0	0	0	0	0	0	0	10	0	0
Blocking and isolation of energies	0	0	0	0	1	1	0	1	0	0	0	0
Burn	0	0	0	0	1	0	0	0	0	0	0	0
Chemical substances	1	1	0	0	11	3	0	1	0	0	0	0
Confined space	0	0	0	0	1	0	0	0	0	0	0	0
Cut	0	0	2	1	7	2	1	1	0	0	0	0
Electrical Shock	0	0	2	0	0	0	0	0	0	0	0	0
Electrical installation	1	0	0	0	0	0	0	0	0	0	0	0
Fall	2	0	1	2	1	1	0	0	0	2	0	0
Fall prevention	0	0	1	0	1	2	1	0	0	1	0	0
Fall prevention (same level)	0	0	0	1	2	4	0	0	0	0	0	0
Individual protection equipment	0	0	1	0	0	0	0	0	0	0	0	0
Liquid Metal	0	0	0	0	2	0	0	1	0	0	0	0
Machine Protection	0	0	0	0	1	1	0	0	0	0	0	0
Manual Tools	1	2	1	1	5	4	0	4	1	1	0	0
Others	41	16	68	43	10	15	8	8	0	15	2	3
Plates	0	0	0	0	0	1	0	0	0	0	0	0
Poll	0	0	0	0	0	0	0	0	0	1	0	0
Power lock	0	0	2	0	0	1	0	0	0	0	0	0
Pressed	2	2	1	2	6	4	0	7	0	0	0	0
Pressurized Systems	0	1	0	0	3	1	0	2	0	0	0	0
Pressurized Systems / Chemical Substances	0	0	0	0	1	2	0	0	0	0	0	0
Projection	3	0	4	0	1	2	2	1	0	0	0	0
Projection of fragments	0	0	0	1	0	0	0	0	0	0	0	0
Projection/Burning	0	0	0	0	1	0	0	0	0	0	0	0
Projection/Choco	0	0	0	0	0	0	0	0	0	1	0	0
Projection/Manual Tools	0	0	0	1	0	0	0	0	0	0	0	0
Suspended Loads	1	0	0	0	3	2	0	0	0	0	0	0
Traffic	0	0	0	0	0	0	0	0	0	0	0	1
Vehicles and Mobile Equipment	4	1	2	1	0	0	0	0	0	0	0	0
Venomous Animals	0	0	0	0	1	0	1	0	1	10	0	0
remains of choco	0	0	4	2	0	0	1	0	0	0	0	0

Critical Risk of type "Others" is dominant across all Locals

5. Local Vs Year

sns.countplot(x="Local",hue="Year", data=data)

<AxesSubplot:xlabel='Local', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Year'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Local  and   Year
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Year
2016	38	13	63	38	45	33	7	18	1	23	2	2
2017	18	10	26	17	14	13	7	9	1	18	0	2

Year 2016 has more accidents across all Local regions compared to 2017

Accident Level Vs Rest All

1. Accident Level Vs Potential Accident Level

sns.countplot(x="Accident Level",hue="Potential Accident Level", data=data)

<AxesSubplot:xlabel='Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Accident Level'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Potential Accident Level')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Potential Accident Level
------------------------------------------------------

Accident Level	I	II	III	IV	V
Potential Accident Level
I	45	0	0	0	0
II	88	7	0	0	0
III	89	14	3	0	0
IV	78	16	26	21	0
V	9	3	2	9	7
VI	0	0	0	0	1

Accident Level I is more related to Potential Accident levels of I, II, III, IV, V, VI

2. Accident Level Vs Natureofemployee

sns.countplot(x="Accident Level",hue="Natureofemployee", data=data)

<AxesSubplot:xlabel='Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Accident Level'],columns=data['Natureofemployee'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Natureofemployee
------------------------------------------------------

Natureofemployee	Employee	Third Party	Third Party (Remote)
Accident Level
I	139	130	40
II	15	19	6
III	14	14	3
IV	10	16	4
V	0	6	2

Accident Level I is more dominant across all Employee types, where Level V is least across all types

3. Accident Level Vs Critical Risk

sns.countplot(x="Accident Level",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a5088c3400>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Accident Level'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Critical Risk')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Critical Risk
------------------------------------------------------

Accident Level	I	II	III	IV	V
Critical Risk
\nNot applicable	0	0	0	1	0
Bees	10	0	0	0	0
Blocking and isolation of energies	3	0	0	0	0
Burn	0	0	1	0	0
Chemical substances	15	2	0	0	0
Confined space	1	0	0	0	0
Cut	11	2	1	0	0
Electrical Shock	2	0	0	0	0
Electrical installation	0	0	0	1	0
Fall	6	0	0	2	1
Fall prevention	5	0	0	1	0
Fall prevention (same level)	6	0	0	1	0
Individual protection equipment	0	1	0	0	0
Liquid Metal	3	0	0	0	0
Machine Protection	2	0	0	0	0
Manual Tools	12	5	3	0	0
Others	169	21	23	13	3
Plates	1	0	0	0	0
Poll	0	0	0	1	0
Power lock	0	0	0	1	2
Pressed	17	1	2	4	0
Pressurized Systems	6	1	0	0	0
Pressurized Systems / Chemical Substances	2	1	0	0	0
Projection	10	2	0	1	0
Projection of fragments	1	0	0	0	0
Projection/Burning	0	1	0	0	0
Projection/Choco	1	0	0	0	0
Projection/Manual Tools	1	0	0	0	0
Suspended Loads	4	0	1	1	0
Traffic	1	0	0	0	0
Vehicles and Mobile Equipment	5	1	0	1	1
Venomous Animals	13	0	0	0	0
remains of choco	2	2	0	2	1

Accident Level I is more domaint with Other critical Risk type

4. Accident Level Vs Year

sns.countplot(x="Accident Level",hue="Year", data=data)

<AxesSubplot:xlabel='Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Year'],columns=data['Accident Level'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Year
------------------------------------------------------

Accident Level	I	II	III	IV	V
Year
2016	211	26	24	19	3
2017	98	14	7	11	5

Accident Level I is more dominant in across 2016 and 2017 years, and Level V is minimum

5. Accident Level Vs Month

sns.countplot(x="Month",hue="Accident Level", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a508b1de80>

bivariate_analysis_df = pd.crosstab(index=data['Accident Level'],columns=data['Month'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Month')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Month
------------------------------------------------------

Month	1	2	3	4	5	6	7	8	9	10	11	12
Accident Level
I	32	42	34	43	31	41	16	15	18	11	9	17
II	2	9	7	2	3	3	1	3	3	4	1	2
III	2	4	3	3	1	2	4	2	2	4	1	3
IV	2	5	3	3	4	4	2	1	1	2	2	1
V	1	1	3	0	1	1	1	0	0	0	0	0

Accident Level 1 dominates across all Months while Level V is minimum

6. Accident Level Vs Country

sns.countplot(x="Country",hue="Accident Level", data=data)

<AxesSubplot:xlabel='Country', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Accident Level'],columns=data['Country'])

print("\n Cross table Analysis of features: ",'Accident Level',' and  ', 'Country')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Accident Level  and   Country
------------------------------------------------------

Country	Country_01	Country_02	Country_03
Accident Level
I	177	98	34
II	19	19	2
III	21	7	3
IV	23	5	2
V	8	0	0

Accident Level I is more dominant across all Countries, while Accident Level V is least dominant across all countries

Potential Accident Level Vs Rest All(Remaining uncoverd)

1. Potential Accident Level Vs Natureofemployee

sns.countplot(x="Potential Accident Level",hue="Natureofemployee", data=data)

<AxesSubplot:xlabel='Potential Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Natureofemployee'])

print("\n Cross table Analysis of features: ",'Potential Accident Level',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Potential Accident Level  and   Natureofemployee
------------------------------------------------------

Natureofemployee	Employee	Third Party	Third Party (Remote)
Potential Accident Level
I	12	29	4
II	44	37	14
III	53	35	18
IV	58	68	15
V	11	15	4
VI	0	1	0

Potential Accident level IV dominents in ThirdParty, while VI is least dominant in Third Party(Remote) across all

2. Potential Accident Level Vs Critical Risk

sns.countplot(x="Potential Accident Level",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a509f5b790>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Potential Accident Level'])

print("\n Cross table Analysis of features: ",'Potential Accident Level',' and  ', 'Critical Risk')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Potential Accident Level  and   Critical Risk
------------------------------------------------------

Potential Accident Level	I	II	III	IV	V	VI
Critical Risk
\nNot applicable	0	0	0	0	1	0
Bees	10	0	0	0	0	0
Blocking and isolation of energies	0	1	2	0	0	0
Burn	0	0	0	1	0	0
Chemical substances	0	5	8	4	0	0
Confined space	0	0	1	0	0	0
Cut	1	6	5	2	0	0
Electrical Shock	0	0	0	2	0	0
Electrical installation	0	0	0	0	1	0
Fall	1	1	4	2	1	0
Fall prevention	1	0	0	5	0	0
Fall prevention (same level)	1	1	3	2	0	0
Individual protection equipment	0	0	0	1	0	0
Liquid Metal	1	0	0	2	0	0
Machine Protection	0	0	2	0	0	0
Manual Tools	2	5	9	4	0	0
Others	16	60	53	85	15	0
Plates	0	1	0	0	0	0
Poll	0	0	0	1	0	0
Power lock	0	0	0	0	3	0
Pressed	2	5	9	7	1	0
Pressurized Systems	0	2	3	2	0	0
Pressurized Systems / Chemical Substances	0	1	0	2	0	0
Projection	0	2	2	7	2	0
Projection of fragments	0	0	0	1	0	0
Projection/Burning	0	0	0	1	0	0
Projection/Choco	0	1	0	0	0	0
Projection/Manual Tools	0	0	1	0	0	0
Suspended Loads	0	1	0	5	0	0
Traffic	0	1	0	0	0	0
Vehicles and Mobile Equipment	0	0	2	2	4	0
Venomous Animals	10	2	1	0	0	0
remains of choco	0	0	1	3	2	1

Among all Critical Risk with Type as "Others" is dominant across all Potential Accident Levels

plt.figure(figsize=(10,6))
sns.barplot(data['Accident Level'], data['Month'], hue=data['Year'], palette='muted') 

<AxesSubplot:xlabel='Accident Level', ylabel='Month'>

3. Potential Accident Level Vs Year

sns.countplot(x="Potential Accident Level",hue="Year", data=data)

<AxesSubplot:xlabel='Potential Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Year'])

print("\n Cross table Analysis of features: ",'Potential Accident Level',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Potential Accident Level  and   Year
------------------------------------------------------

Year	2016	2017
Potential Accident Level
I	26	19
II	69	26
III	75	31
IV	97	44
V	16	14
VI	0	1

There is Decrease in Number of accidents across all Potential Accident level from 2016 to 2017. Potential Accident level IV is dominant in both 2016 and 2017

4. Potential Accident Level Vs Country

sns.countplot(x="Potential Accident Level",hue="Country", data=data)

<AxesSubplot:xlabel='Potential Accident Level', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Potential Accident Level'],columns=data['Country'])

print("\n Cross table Analysis of features: ",'Potential Accident Level',' and  ', 'Country')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Potential Accident Level  and   Country
------------------------------------------------------

Country	Country_01	Country_02	Country_03
Potential Accident Level
I	10	6	29
II	51	40	4
III	64	41	1
IV	101	33	7
V	21	9	0
VI	1	0	0

Potential Accident Level IV is dominant across all countries, while with VI least number of accidents happenned

Natureofemployee Vs RestAll

1. Natureofemployee Vs Critical Risk

sns.countplot(x="Natureofemployee",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a50a4a3310>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Natureofemployee'])

print("\n Cross table Analysis of features: ",'Critical Risk',' and  ', 'Natureofemployee')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Critical Risk  and   Natureofemployee
------------------------------------------------------

Natureofemployee	Employee	Third Party	Third Party (Remote)
Critical Risk
\nNot applicable	1	0	0
Bees	1	9	0
Blocking and isolation of energies	2	0	1
Burn	1	0	0
Chemical substances	9	2	6
Confined space	1	0	0
Cut	8	5	1
Electrical Shock	0	0	2
Electrical installation	0	1	0
Fall	0	5	4
Fall prevention	2	3	1
Fall prevention (same level)	3	4	0
Individual protection equipment	0	1	0
Liquid Metal	3	0	0
Machine Protection	1	1	0
Manual Tools	7	7	6
Others	99	109	21
Plates	1	0	0
Poll	0	1	0
Power lock	0	0	3
Pressed	12	7	5
Pressurized Systems	4	1	2
Pressurized Systems / Chemical Substances	1	1	1
Projection	7	6	0
Projection of fragments	0	1	0
Projection/Burning	1	0	0
Projection/Choco	0	1	0
Projection/Manual Tools	0	1	0
Suspended Loads	4	1	1
Traffic	1	0	0
Vehicles and Mobile Equipment	3	5	0
Venomous Animals	3	9	1
remains of choco	3	4	0

Critical Risk of type "Others" is dominant across all Types of Employees

Year Vs RestAll

1. Year vs month

sns.countplot(x="Month",hue="Year", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a50a51aca0>

bivariate_analysis_df = pd.crosstab(index=data['Month'],columns=data['Year'])

print("\n Cross table Analysis of features: ",'Month',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Month  and   Year
------------------------------------------------------

Year	2016	2017
Month
1	12	27
2	31	30
3	34	16
4	28	23
5	26	14
6	31	20
7	19	5
8	21	0
9	24	0
10	21	0
11	13	0
12	23	0

From the above plot, it is evident that the max accidents happened in the year 2016 and march.

2. Year vs Local

sns.countplot(x="Local",hue="Year", data=data)

<AxesSubplot:xlabel='Local', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Year'],columns=data['Local'])

print("\n Cross table Analysis of features: ",'Local',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Local  and   Year
------------------------------------------------------

Local	1	2	3	4	5	6	7	8	9	10	11	12
Year
2016	38	13	63	38	45	33	7	18	1	23	2	2
2017	18	10	26	17	14	13	7	9	1	18	0	2

From the above plot, it can be determined that the maximum accidents took place in the local 3 and year 2016.

3. Year vs Weekday

sns.countplot(x="Weekday",hue="Year", data=data)

<AxesSubplot:xlabel='Weekday', ylabel='count'>

bivariate_analysis_df = pd.crosstab(index=data['Year'],columns=data['Weekday'])

print("\n Cross table Analysis of features: ",'Weekday',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Weekday  and   Year
------------------------------------------------------

Weekday	Friday	Monday	Saturday	Sunday	Thursday	Tuesday	Wednesday
Year
2016	44	40	36	25	58	40	40
2017	17	13	20	16	18	29	22

From the above plot, it is clearly evident that maximum number of accidents took place on thursday and year 2016.

4.Year vs Critical Risk

sns.countplot(x="Year",hue="Critical Risk", data=data)
plt.legend(bbox_to_anchor=(1.02, 1), loc='upper left', borderaxespad=0)

<matplotlib.legend.Legend at 0x1a50aa04b20>

bivariate_analysis_df = pd.crosstab(index=data['Critical Risk'],columns=data['Year'])

print("\n Cross table Analysis of features: ",'Critical Risk',' and  ', 'Year')
print("------------------------------------------------------")
display(bivariate_analysis_df)

 Cross table Analysis of features:  Critical Risk  and   Year
------------------------------------------------------

Year	2016	2017
Critical Risk
\nNot applicable	1	0
Bees	10	0
Blocking and isolation of energies	3	0
Burn	0	1
Chemical substances	13	4
Confined space	1	0
Cut	6	8
Electrical Shock	0	2
Electrical installation	1	0
Fall	2	7
Fall prevention	1	5
Fall prevention (same level)	6	1
Individual protection equipment	0	1
Liquid Metal	2	1
Machine Protection	0	2
Manual Tools	14	6
Others	189	40
Plates	1	0
Poll	1	0
Power lock	0	3
Pressed	14	10
Pressurized Systems	7	0
Pressurized Systems / Chemical Substances	3	0
Projection	1	12
Projection of fragments	0	1
Projection/Burning	0	1
Projection/Choco	0	1
Projection/Manual Tools	0	1
Suspended Loads	5	1
Traffic	1	0
Vehicles and Mobile Equipment	0	8
Venomous Animals	1	12
remains of choco	0	7

From the above plot, it is clearly evident that maximum number of accidents took place with "Others" and year 2016.

Multivariate analysis

#Pair plot
sns.pairplot(data)

<seaborn.axisgrid.PairGrid at 0x1a50aa458b0>

data.corr()

	Local	Year	Month
Local	1.000000	0.054246	0.019061
Year	0.054246	1.000000	-0.416621
Month	0.019061	-0.416621	1.000000

From the above Correlation diagram its clear that "Local" and "Year" are moderately correlated

data.columns

Index(['Date', 'Country', 'Local', 'Industry Sector', 'Accident Level',
       'Potential Accident Level', 'Gender', 'Natureofemployee',
       'Critical Risk', 'Description', 'Year', 'Month', 'Weekday', 'Season'],
      dtype='object')

**Groupby Analysis**

Year wise distribution of accidents and potential accident levels

data.groupby(['Year','Accident Level','Potential Accident Level'])[['Accident Level']].count()

			Accident Level
Year	Accident Level	Potential Accident Level
2016	I	I	26
		II	62
		III	64
		IV	53
		V	6
	II	II	7
		III	9
		IV	10
	III	III	2
		IV	20
		V	2
	IV	IV	14
		V	5
	V	V	3
2017	I	I	19
		II	26
		III	25
		IV	25
		V	3
	II	III	5
		IV	6
		V	3
	III	III	1
		IV	6
	IV	IV	7
		V	4
	V	V	4
		VI	1

1. Year 2016 with Accident Level I has maximum accidents of 64 with Potential Accident Level III and 62 with Potential Accident Level II

2. Year 2017 with Accident Level I has maximum accidents of 26 with Potential Accident Level II and 25 with Potential Accident Level III,IV

Year wise distribution of Industry Sector and accident levels

data.groupby(['Year','Industry Sector','Accident Level'])[['Accident Level']].count()

			Accident Level
Year	Industry Sector	Accident Level
2016	Metals	I	79
		II	9
		III	4
		IV	5
	Mining	I	112
		II	15
		III	17
		IV	12
		V	3
	Others	I	20
		II	2
		III	3
		IV	2
2017	Metals	I	28
		II	3
		III	3
		IV	2
		V	1
	Mining	I	51
		II	11
		III	3
		IV	9
		V	4
	Others	I	19
		III	1

1. Year 2016 with Industry Sector of Type "Metals" has maximum accidents of 79 with Accident Level I
2. Year 2016 with Industry Sector of Type "Mining" has maximum accidents of 112 with Accident Level I

3. Year 2016 with Industry Sector of Type "Others" has maximum accidents of 20 with Accident Level I

4. Year 2017 with Industry Sector of Type "Metals" has maximum accidents of 28 with Accident Level I

5. Year 2017 with Industry Sector of Type "Mining" has maximum accidents of 51 with Accident Level I
6. Year 2017 with Industry Sector of Type "Others" has maximum accidents of 19 with Accident Level I

Industry Sector wise distribution of Country and accident levels

data.groupby(['Industry Sector','Country','Accident Level'])[['Accident Level']].count()

			Accident Level
Industry Sector	Country	Accident Level
Metals	Country_01	I	36
		II	1
		III	3
		IV	5
		V	1
	Country_02	I	71
		II	11
		III	4
		IV	2
Mining	Country_01	I	140
		II	18
		III	17
		IV	18
		V	7
	Country_02	I	23
		II	8
		III	3
		IV	3
Others	Country_01	I	1
		III	1
	Country_02	I	4
	Country_03	I	34
		II	2
		III	3
		IV	2

1. Metals in Country_01 has maximum accidents with Level 1 with 36 count
2. Metals in Country_02 has maximum accidents with Level 1 with 71 count
3. Mining in Country_01 has maximum accidents with Level 1 with 140 count
4. Mining in Country_02 has maximum accidents with Level 1 with 23 count
5. Others in Country_03 has maximum accidents with Level 1 with 34 count

Word Cloud Analysis

#!pip install wordcloud
#!pip install pandas_profiling
from wordcloud import WordCloud

WordCloud for Accident Level and Description

for i in data['Accident Level'].unique():
    print('WordCloud for Accident Level :', i,'\n')
    text = " ".join(cat.split()[1] for cat in data[data['Accident Level'] == i]['Description'])
    # Creating word_cloud with text as argument in .generate() method
    word_cloud = WordCloud(collocations = False, background_color = 'lightyellow').generate(text)
    # Display the generated Word Cloud
    plt.figure(figsize=[10,10])
    
    plt.imshow(word_cloud, interpolation='bilinear')
    plt.axis("off")
    plt.show
    print('-----------------------------')
    

WordCloud for Accident Level : I 

-----------------------------
WordCloud for Accident Level : IV 

-----------------------------
WordCloud for Accident Level : III 

-----------------------------
WordCloud for Accident Level : II 

-----------------------------
WordCloud for Accident Level : V 

-----------------------------

WordCloud for Potential Accident Level and Description

for i in data['Potential Accident Level'].unique():
    if i != 'VI':
        print('WordCloud for Potential Accident Level :', str(i),'\n')
        text = " ".join(cat.split()[1] for cat in data[data['Potential Accident Level'] == i]['Description'])
        # Creating word_cloud with text as argument in .generate() method
        word_cloud = WordCloud(collocations = False, background_color = 'lightyellow').generate(text)
        # Display the generated Word Cloud
        plt.figure(figsize=[10,10])

        plt.imshow(word_cloud, interpolation='bilinear')
        plt.axis("off")
        plt.show
        print('-----------------------------')

WordCloud for Potential Accident Level : IV 

-----------------------------
WordCloud for Potential Accident Level : III 

-----------------------------
WordCloud for Potential Accident Level : I 

-----------------------------
WordCloud for Potential Accident Level : II 

-----------------------------
WordCloud for Potential Accident Level : V 

-----------------------------

WordCloud for Industry Sector and Description

for i in data['Industry Sector'].unique():
    print('WordCloud for Industry type :', str(i),'\n')
    text = " ".join(cat.split()[1] for cat in data[data['Industry Sector'] == i]['Description'])
    # Creating word_cloud with text as argument in .generate() method
    word_cloud = WordCloud(collocations = False, background_color = 'lightyellow').generate(text)
    # Display the generated Word Cloud
    plt.figure(figsize=[10,10])
    
    plt.imshow(word_cloud, interpolation='bilinear')
    plt.axis("off")
    plt.show
    print('-----------------------------')

WordCloud for Industry type : Mining 

-----------------------------
WordCloud for Industry type : Metals 

-----------------------------
WordCloud for Industry type : Others 

-----------------------------

WordCloud for Country and Description

for i in data['Country'].unique():
    print('WordCloud for Country :', i,'\n')
    text = " ".join(cat.split()[1] for cat in data[data['Country'] == i]['Description'])
    # Creating word_cloud with text as argument in .generate() method
    word_cloud = WordCloud(collocations = False, background_color = 'lightyellow').generate(text)
    # Display the generated Word Cloud
    plt.figure(figsize=[10,10])
    
    plt.imshow(word_cloud, interpolation='bilinear')
    plt.axis("off")
    plt.show
    print('-----------------------------')
    

WordCloud for Country : Country_01 

-----------------------------
WordCloud for Country : Country_02 

-----------------------------
WordCloud for Country : Country_03 

-----------------------------

from pandas_profiling import ProfileReport
profile = ProfileReport(data, title="Pandas Profiling Report")
profile.to_notebook_iframe()